← All reports

Claude can be used as a tool to facilitate hacking into other AI companies' infrastructure.

AICybersecuritySep 18, 2026score 0.172 posts · 0 replies across 1 instances
The posts report on a security research finding where researchers utilized the AI model Claude to successfully breach OpenAI's systems. This highlights potential vulnerabilities in AI security and the risks of using LLMs for cyberattacks.

Claims

Claude can be used as a tool to facilitate hacking into other AI companies' infrastructure.
Parent: AIEntity: ClaudeSub-entity: security researchImpact: negativeDate: Sep 18, 2026Target: Claude's security utility

Source posts

@[email protected]
#ArsTechnica #Tech Researchers used Claude to hack OpenAI https://arstechni.ca/tSGm
0 boosts · 0 favs · 0 replies · Sep 18, 2026
#arstechnica#tech
@[email protected]
#TheVerge #Tech Security researchers used Claude to help them hack into OpenAI https://thever.ge/DYA4
0 boosts · 0 favs · 0 replies · Sep 18, 2026
#theverge#tech